Skip to main content
NCBI home page
HHS Author Manuscripts logoLink to HHS Author Manuscripts
. Author manuscript; available in PMC: 2024 Sep 13.
Published in final edited form as: Regul Toxicol Pharmacol. 2023 May 18;141:105410. doi: 10.1016/j.yrtph.2023.105410

Revisiting the mutagenicity and genotoxicity of N-nitroso propranolol in bacterial and human in vitro assays

Xilin Li a,**, Yuan Le a, Ji-Eun Seo a, Xiaoqing Guo a, Yuxi Li a, Si Chen a, Roberta A Mittelstaedt a,1, Nyosha Moore a,1, Sharon Guerrero a, Audrey Sims a, Sruthi T King b, Aisar H Atrakchi b, Timothy J McGovern b, Karen L Davis-Bruno b, David A Keire b, Rosalie K Elespuru c, Robert H Heflich a, Nan Mei a,*
PMCID: PMC11393638  NIHMSID: NIHMS2019427  PMID: 37210026

Abstract

Propranolol is a widely used β-blocker that can generate a nitrosated derivative, N-nitroso propranolol (NNP). NNP has been reported to be negative in the bacterial reverse mutation test (the Ames test) but genotoxic in other in vitro assays. In the current study, we systematically examined the in vitro mutagenicity and genotoxicity of NNP using several modifications of the Ames test known to affect the mutagenicity of nitrosamines, as well as a battery of genotoxicity tests using human cells. We found that NNP induced concentration-dependent mutations in the Ames test, both in two tester strains that detect base pair substitutions, TA1535 and TA100, as well as in the TA98 frameshift-detector strain. Although positive results were seen with rat liver S9, the hamster liver S9 fraction was more effective in bio-transforming NNP into a reactive mutagen. NNP also induced micronuclei and gene mutations in human lymphoblastoid TK6 cells in the presence of hamster liver S9. Using a panel of TK6 cell lines that each expresses a different human cytochrome P450 (CYP), CYP2C19 was identified as the most active enzyme in the bioactivation of NNP to a genotoxicant among those tested. NNP also induced concentration-dependent DNA strand breakage in metabolically competent 2-dimensional (2D) and 3D cultures of human HepaRG cells. This study indicates that NNP is genotoxic in a variety of bacterial and mammalian systems. Thus, NNP is a mutagenic and genotoxic nitrosamine and a potential human carcinogen.

Keywords: Nitrosamine impurities, N-Nitroso propranolol, Cytochrome P450s, Hamster liver S9, Gene mutation, Chromosomal damage

1. Introduction

Propranolol hydrochloride is a widely used β-adrenergic receptor blocker (β-blocker) that is used for the treatment of hypertension, angina, heart rhythm disorders, and other cardiovascular conditions (Kalam et al., 2020). Propranolol itself appears to be relatively safe, even in infants (Barton et al., 2015). However, under acidic conditions, it can potentially react with nitrite and produce N-nitroso propranolol (NNP) (Fig. 1), raising safety concerns (Sluggett et al., 2018). Currently, all N-nitroso compounds are considered a “cohort-of-concern” in the ICH M7(R1) guideline (FDA, 2018; ICH, 2017), because most of the compounds in this class that have been tested are carcinogenic in rodent bioassays (Lijinsky and Epstein, 1970; Magee and Barnes, 1967; Preussmann, 1984). Therefore, the presence of NNP in drug products could increase the risk of cancer in subjects using propranolol when exposed above acceptable levels and over long periods of time, and propranolol batches have been recalled in some regions due to the presence of NNP (Canada, 2022). Due to the wide use of propranolol, the U.S. Food and Drug Administration undertook an investigation to evaluate the risk of NNP and to identify optimized testing conditions for its safety assessment in vitro.

Fig. 1.

Fig. 1.

Open in a new tab

Nitrosation of propranolol under acidic condition produces N-nitroso propranolol.

The available evidence on the mutagenicity and genotoxicity of NNP is inconsistent. We have identified only two publications describing data on the in vitro mutagenicity of NNP that employed the Ames test and a mammalian cell assay in Chinese hamster V79 cells (Raisfeld-Danse and Chen, 1983; Zhang et al., 1983). One study showed that 1 mg/plate NNP did not induce mutations in tester strains TA98 or TA100 in the presence of Aroclor 1254-induced rat liver S9 (Zhang et al., 1983). The second study tested NNP at a maximum concentration of 1.4 μmol/plate (404 μg/plate) with tester strains TA1535, TA1537, TA1538, TA92, TA98, and TA100, and found that NNP was negative with and without Aroclor 1254-induced rat liver S9 (Raisfeld-Danse and Chen, 1983). Two additional studies were conducted using the nitrosation products of propranolol. While not testing NNP itself, these studies found that the propranolol-nitrite reaction products were mutagenic in tester strains TA1535 and TA98 (Kikugawa et al., 1987; Ozhan and Alpertunga, 2003).

In contrast to the mixed findings on its mutagenicity, NNP was shown to be genotoxic in primary rat and human hepatocytes. At subtoxic concentrations ranging from 10 μM to 100 μM, 20-h exposure to NNP induced a concentration-dependent increase in DNA fragmentation and unscheduled DNA synthesis (UDS) in hepatocytes (Robbiano et al., 1991). In another experiment, a single dose of 1000 mg/kg NNP induced micronucleus (MN) formation in the hepatocytes of Sprague-Dawley rats, a potential indicator of clastogenic effects (Martelli et al., 1994).

In general, the mutagenicity of most N-nitroso compounds is dependent upon metabolic activation. Phase I drug metabolizing enzymes such as cytochrome P450 (CYP) 2E1, CYP2A6, CYP2C9, CYP3A4, and CYP1A1 bioactivate nitroso dialkylamines of different sizes and structures, largely via hydroxylation on the α-carbon (carbon adjacent to the N-nitroso moiety) (Bellec et al., 1996; Fujita and Kamataki, 2001; Yamazaki et al., 1992). For nitrosamine drug substance-related impurities (NDSRIs), a term that refers to nitrosamine impurities sharing structural similarity to the active pharmaceutical ingredient (nitroso group added to the secondary or tertiary amine), the necessity of metabolic activation for manifesting their genotoxicity may depend on the structure among other factors. It appears that NNP, along with other nitroso derivatives generated from β-blockers, requires metabolic activation to exert its genotoxic effects (Robbiano et al., 1991). NNP has α-carbon atoms with hydrogens available for hydroxylation; however, questions remain as to whether such metabolism occurs for NNP, what specific drug metabolizing enzymes are involved in its biotransformation, and whether any metabolites of NNP can produce mutations.

In the current study, we systematically examined the in vitro mutagenicity and genotoxicity of NNP. The Ames test was used to evaluate the mutagenicity of NNP in bacteria under different test conditions known to affect the mutagenicity of nitrosamines. In addition, several assays were conducted to evaluate the genotoxicity of NNP in mammalian cells. We used human lymphoblastoid TK6 cells with hamster liver S9 as an exogenous metabolic activation system. To identify any specific human enzymes involved in the bioactivation of NNP, we also conducted assays using a panel of TK6 cell lines each of which stably expresses a specific human CYP (Li et al. 2020a, 2020b). Lastly, the genotoxicity of NNP was evaluated using metabolically competent human HepaRG cells, which possess a variety of drug metabolizing enzymes found in primary human hepatocytes (Seo et al., 2019).

2. Materials and methods

2.1. Chemicals

NNP (CAS# 84418-35-9) was custom synthesized by Clearsynth (Ontario, Canda) with a purity of 99.69%. Positive control chemicals, including 2-aminoanthracene (2-AA, CAS# 613-13-8), acridine mutagen ICR 191 (CAS# 17070-45-0), 2-nitrofluorene (2-NF, CAS# 607-57-8), methyl methanesulfonate (MMS, CAS# 66-27-3), mitomycin C (CAS# 50-07-7), and N-nitrosodiethylamine (NDEA, CAS# 55-18-5) were purchased from Sigma Aldrich (St. Louis, MO). N-Nitrosodimethylamine (NDMA, CAS# 62-75-9) was purchased from Chem Service (West Chester, PA) and 1-cyclopentyl-4-nitroso-piperazine (CPNP, CAS# 61379-66-6) from Toronto Research Chemicals (Toronto, Canada). All chemicals were stored as recommended by the vendors. Chemical solutions were freshly prepared before treatments.

2.2. Ames tester strains and preparation

The tester strains used were the Salmonella typhimurium histidine auxotrophs TA98, TA100, TA1535 and TA1537, and the E. coli tryptophan auxotroph, WP2 uvrA (pKM101). All tester stains were obtained from MOLTOX (Boone, NC) and stored either on paper disks at 4 °C or as frozen stocks at −80 °C. Overnight cultures were prepared by inoculating 0.1 mL of frozen stock cultures or a disk infused with the tester strain into a vessel containing 10 mL of culture medium (Oxoid Nutrient Broth #2). To assure that cultures were harvested in late log phase, the length of incubation was controlled and monitored. Following inoculation, each flask was placed in a shaker/incubator programmed to begin shaking at 200 rpm and incubating at 37 °C approximately 14 h before the anticipated time of harvest. Each culture was monitored spectrophotometrically for turbidity and was harvested at a percent transmittance yielding a titer equal to or greater than 0.3 × 109 cells/mL.

2.3. Human TK6, HepG2, and HepaRG cell cultures

Human lymphoblastoid TK6 cells were purchased from American Type Culture Collection (ATCC; Manassas, VA). Previously, we generated eight TK6 cell lines individually expressing human CYP1A1, 2A6, 2C8, 2C9, 2C19, 2D6, 2E1, and 3A4 (Li et al., 2020a, 2020b). The TK6 cell line and CYP-expressing TK6 cell lines were cultured at 37 °C in a humidified atmosphere of 5% CO2 in air, using RPMI 1640 medium supplemented with L-glutamine (Thermo Fisher Scientific, Waltham, MA), 100 U/mL penicillin (Thermo Fisher), 100 μg/mL streptomycin (Thermo Fisher), and 10% horse serum (Thermo Fisher). Cells were routinely maintained at a density of 2 × 105 to 1.5 × 106 cells/mL.

Human HepG2 cells were obtained from ATCC (Manassas, VA). HepG2 cells transduced with human CYP2C19 were generated previously (Chen et al., 2021). Cells were cultured in Williams’ Medium E complete media containing 10% fetal bovine serum (FBS) and antibiotics. HepG2 cells and CYP2C19-expressing HepG2 cells were seeded at a concentration of 2.5 × 105 cells/mL in volumes of 10 mL in 100-mm tissue culture plates. Cells were routinely maintained at a confluency between 20% and 80%.

The HepaRG human hepatoma cell line was obtained from Biopredic International (Saint Grégoire, France) and cultured as previously described (Seo et al., 2022). Briefly, 0.65–0.7 × 106 undifferentiated HepaRG cells were seeded in a 100-mm tissue culture dish and cultured in growth medium for 14 days. The cells were then cultured in differentiation medium for another 14 days. The growth and differentiation media were prepared by adding growth and differentiation additives (Lonza, Walkersville, MD), respectively, to William’s E Medium (Thermo Fisher) supplemented with 2 mM GlutaMax (Thermo Fisher) and 100 μg/mL primocin (InvivoGen, San Diego, CA). Fully differentiated HepaRG cells were dissociated with TrypLE Express (Thermo Fisher) and re-seeded into a 96-well flat bottom plate (Corning Inc., Corning, NY) at a density of 5 × 104 cells/well or a 384-well ultra-low attachment (ULA) plate (Corning) at a density of 5 × 103 cells/well to form 2-dimensional (2D) monolayer cultures or 3D spheroids, respectively. After incubation for three days (for 2D cultures) or 10 days (for 3D spheroids), cells were exposed to various concentrations of NNP (dissolved in dimethyl sulfoxide, DMSO) for 24 h at 37 °C in a humidified atmosphere of 5% CO2 in air.

2.4. Exogenous metabolic activation

A previous study (Lijinsky and Andrews, 1983) and our preliminary results showed that both Aroclor 1254-induced rat and hamster liver S9s have the ability to activate NDMA (and other N-nitrosamines) to a mutagen in the Ames test, with hamster S9 producing more robust responses. As Aroclor 1254 is no longer available as an enzyme inducer for animals used in S9 preparations, the liver S9s used in the current study were prepared from male Sprague-Dawley rats or male Syrian hamsters pretreated with a combination of phenobarbital and β-naphthyl flavone (PB/BNF), which has been shown to be comparable to Aroclor 1254 as an enzyme inducer (Callander et al., 1995). The S9s were purchased from MOLTOX (Boone, NC), stored at −80 °C, and thawed immediately before use.

2.5. Bacterial reverse mutation test

The Ames test was conducted following Organization for Economic Co-operation and Development (OECD) Test Guideline (TG) 471 (OECD, 2020). As recommended for nitrosamines, a precubation step was performed by mixing liver S9, cofactors, bacterial tester strain, and the test compound and subsequently incubating the mixture at 37 °C and 300 rpm for 30 or 60 min prior to plating the contents of the tube onto soft agar. A total of five metabolic activation conditions, including 0%, 10% or 30% (v/v) rat liver S9, and 10% or 30% hamster liver S9, were used to evaluate the mutagenicity of NNP (dissolved in acetone). The cofactors in the preincubation included 5 mM glucose-6-phosphate, 4 mM NADP, 33 mM KCl, 8 mM MgCl2, and 100 mM phosphate-buffered saline (PBS, pH 7.4). A preliminary study showed NNP had considerable toxicity at concentrations of 100–500 μg/plate and above. Therefore, the definitive mutagenicity assays were conducted at NNP concentrations of 0 (vehicle control), 10, 25, 50, 100 and 200 μg per plate. Positive control chemicals were included for each tester strain as indicated in the Supplementary Materials.

Preincubations were conducted in triplicate in blunted-bottom (non-conical) 2-mL polypropylene microcentrifuge tubes. Following preincubation for 30 or 60 min, the contents of individual tubes were transferred into Fisherbrand 13 × 100 mm borosilicate glass tubes containing 2 mL molten, 45 °C top agar (0.6% agar, 0.45% NaCl, with 0.05 mM tryptophan for the E. coli strain or 0.05 mM histidine and biotin for the Salmonella strains). The suspension was mixed gently and poured onto 100-mm Minimal Glucose Agar plates (MOLTOX). The plates were gently swirled, the top agar allowed to solidify, and then the plates were inverted, agar-side-up and incubated for 2 or 3 days at 37 °C. Revertant colonies were enumerated with a ProtoCOL2 automated plate counter (Synbiosis USA, Frederick, MD). For strains TA1535 and TA1537, results were judged positive if the maximum mean revertants/plate induced by NNP was equal to or greater than 3-fold over the mean solvent control value. For strains TA98, TA100, and WP2 uvrA (pKM101), results were judged positive if maximum mean revertants/plate induced by NNP was equal to or greater than 2-fold over the mean solvent control value.

Benchmark dose (BMD) analysis was conducted using web-based PROAST software (version 70.0, available at https://r4eu.efsa.europa.eu/). The covariate approach and a Critical Effect Size (CES) of 0.5 (50% increase relative to the vehicle control) were used on dose-response curves fit set of models. Model averaging was used to estimate BMD values. The number of bootstrap runs for calculating model-averaged BMD confidence invervals was 200. Analyses were constructed to evaluate the effect of the type of metabolic activation and time of preincubation on the extent of mutagenicity. Differences were based on the occurrence of non-overlapping BMD confidence intervals for the dose responses (Mittelstaedt et al., 2021; Wills et al., 2016).

2.6. In vitro micronucleus assay

For the flow cytometric analysis of the in vitro MN assay (Bryce et al., 2007), cells were stained and lysed following the protocol described in the In Vitro MicroFlow Kit (Litron Laboratories, Rochester, NY). Ethidium monoazide (EMA) and SYTOX Green were used to stain apoptotic/necrotic cells and chromatin, respectively. Cells were analyzed using a BD FACSCanto II Cell Analyzer equipped with a highthroughput sampler (BD Biosciences, San Jose, CA). The stopping gate was set to record 10,000 intact nuclei.

TK6 cells were seeded in 24-well cell culture plates at a concentration of 2 × 105 cells/mL and exposed to vehicle (DMSO), NNP, or positive controls continuously for 4 h with S9 and for 24 h with or without S9. Following the 4-h treatment, the cells were cultured in fresh medium for an additional 20 h to allow the cells to go through 1.5–2 cell cycles, as recommended by OECD TG 487 (OECD, 2016). The S9 mix protocol was based on previous studies (Avlasevich et al., 2021; O’Neill et al., 1982). Briefly, PB/BNF-induced hamster liver S9 mix was prepared containing S9 protein, 50 mM sodium phosphate (pH 8.0), 5 mM glucose-6-phosphate, 4 mM NADP, 30 mM KCl, 10 mM MgCl2, and 10 mM CaCl2, with the CaCl2 added last. The addition of calcium cations has been shown to increase the sensitivity for detecting the genotoxicity of nitrosamines in mammalian cells (O’Neill et al., 1982). The S9 concentrations for the 4-h and 24-h treatments were 10% and 2% (v/v), respectively; and subsequently, one volume of the S9 mix was added into four volumes of cell growth medium to make a final S9 concentration of 2% and 0.4%, for the 4-h and 24-h treaments, repsectively. After the treatment, the cells were washed three times with warm PBS, which helped eliminate the calcium phosphate precipitate and other components of the S9 mix. To identify the specific human enzymes accounting for the bioactivation of NNP, a panel of TK6 cells, each endogenously expressing a human CYP, were seeded at 2 × 105 cells/mL and treated with vehicle or NNP continuously for 24 h (without S9).

For the HepG2 cell MN test, cells were seeded at a density of 2.5 × 105 cells/mL in a volume of 100 μL per well in 96-well plates and treated with vehicle or NNP continuously for 24 h (without S9). Subsequently, cells were cultured for additional 16 h in fresh medium to achieve a 1.5- to 2.0-fold cell population increase before harvest.

2.7. In vitro Comet assay

NNP-induced DNA strand breaks were evaluated in 2D and 3D HepaRG models using the CometChip assay with concurrent cytotoxicity assessment by the ATP assay as previously described (Seo et al., 2022). Following a 24-h treatment, 2D HepaRG cells and 3D spheroids were dissociated into single cell suspensions. Subsequently, the cells were loaded by gravity into each microwell of a 96-well CometChip (Trevigen, Gaithersburg, MD). Then the cells were lysised, followed by electrophoresis under alkaline conditions to measure single- and double-strand breaks. Comet images were acquired automatically using a Cytation 5 Cell Imaging MultiMode Reader (BioTek, Winooski, VT) and the percent of tail DNA was determined using Trevigen Comet Analysis Software. A relative cell viability of at least 70% measured by the ATP assay was used as a cut-off value to minimize false positive responses in the Comet assay.

2.8. In vitro MultiFlow DNA damage assay

The MultiFlow flow-cytometry-based assay (Litron Laboratories) was used to measure multiple endpoints associated with DNA damage response pathways. TK6 cells or CYP2C19-expressing TK6 cells were seeded in 24-well plates at a concentration of 2 × 105 cells/mL and exposed to NNP, vehicle control (DMSO), or positive controls in the presence of hamster liver S9. After a 4-h treatment with a 20-h recovery peroid or 24 h of continous treatment, the cells were resuspended in PBS and 25 μL of the cell suspension were added to the wells of a separate 96-well round bottom plate containing 50 μL/well of complete labeling solution prepared following the manufacturer’s instructions. The labeling solution contained anti-γH2A.X-Alexa Fluor 647, anti-phospho-histone H3-PE, and anti-P53-FITC antibodies to detect DNA double-strand breaks (DSBs) and mitotic cells. After a 1-h incubation at room temperature, the cells were analyzed using a BD FACSCanto II Cell Analyzer equipped with a highthroughput sampler. The flow cytometric gating and data analysis employed methods described previously (Bryce et al., 2016).

2.9. In vitro gene mutation assays

To measure forward mutations in reporter genes such as thymidine kinase (TK) and hypoxanthine-guanine phosphoribosyltransferase (HPRT) genes, the in vitro mammalian cell gene mutation assays were conducted following OECD TGs 476 and 490 with minor modifications (OECD, 2016). Before the treatments, TK6 cells and CYP2C19-expressing TK6 cells were cleansed for two days by growth in CHAT medium (10 μM 2′-deoxycytidine, 200 μM hypoxanthine, 0.1 μM aminopterin, and 17.5 μM thymidine) containing 10% horse serum, and then cultured in HCT medium (CHAT medium without aminopterin) for another two days (Chen et al., 2022). For TK6 cells only, PB/BNF-induced hamster liver S9 mix containing 5% S9 protein (v/v) and cofactors was prepared for exogenous metabolic activation; after dilution into the treatment medium containing 5% horse serum, the final S9 concentration was 1%. For the exposure of the cells to vehicle control (DMSO), NNP, and positive controls at the indicated concentrations, 10 mL cleansed cells were ali-quoted at a concentration of 1 × 106 cells/mL, treated with the chemicals in the presence (for TK6 cells) or absence (for CYP2C19-expressing TK6 cells) of S9 mix, and then incubated in a roller drum at 37 °C for 4 h. After chemical treatments, the cells were washed twice with 1 × PBS and re-suspended in the culture medium containing 10% horse serum. To determine the cloning efficiency immediately following the treatments (CE0), the cells of each treatment were plated in two 96-microwell plates at a concentration of 8 cells/mL (about 1.6 cell/200 μL/well) and colonies were scored after 14 days. Cytotoxicity for each treatment was calculated from CE0 and determined as the relative survival (RS%) of the vehicle control.

Treated cells were maintained at a density of 3–4 × 105 cells/mL in a 25 mL volume and cultured for 3 days or 7 days allowing for the expression of the TK- or HPRT-deficient phenotype, respectively. Thereafter, cells were plated at 40,000 cells/200 μL/well in four 96-well plates with the selective medium containing 20% horse serum and 4 μg/mL triflurothymidine (TFT) or 3 μg/mL 6-thioguanine (6-TG) for TK or HPRT mutant selection, respectively. At the same time, the cloning efficiency of the cell population, as part of TK mutant analysis on the 3rd day (CE3) and HPRT mutant analysis on the 7th day (CE7), was determined by plating cells suspended in the non-selective medium (i.e., the selective medium without TFT or 6-TG) into two 96-well plates at a mean density of 1.6 cell/200 μL/well. After 14 days of incubation, the numbers of colonies in the CE3 and TFT selection plates (i.e., the normal growing colonies or early appearing TK mutants) were scored. Following this, the TK selection plates were re-fed with 100 μL/well of fresh medium containing TFT (12 μg/mL) and incubated for an additional 7 days; the slow growing colonies (i.e., late appearing/slow-growth TK mutants) were then counted after 21 days of incubation. In addition, the numbers of colonies in the CE7 and 6-TG selection plates were counted after 14 days to determine HPRT mutant frequencies (MFs). For data analysis, the cloning efficiency of cells in both selective medium and non-selective medium was determined according to the Poisson distribution, and the TK or HPRT MF for each treatment was calculated from these two cloning efficiencies.

2.10. Statistical analysis

All data from the in vitro MultiFlow DNA damage assay, Comet assay, and MN assay are presented as mean ± standard deviation (SD) of data from at least three independent experiments. Statistical analyses were performed using GraphPad Prism version 6.0 (GraphPad Software, La Jolla, CA). One-way ANOVA followed by Dunnett’s post hoc test was used to evaluate differences between groups, with statistical significance set at P < 0.05.

3. Results

3.1. Mutagenicity in the ames test

The Ames test was conducted over a dose-range of 0–200 μg NNP/plate using the preincubation method. Higher concentrations were not tested due to overt cytotoxicity. As summarized in Table 1 (detailed data in Supplementary Tables 1–10), positive mutagenic responses were observed in the tester strains detecting base pair substitution at G:C base pairs (TA1535 and TA100) and in the TA98 frameshift detector strain (also having a GC target sequence). S9 activation was required for the mutagenic responses in TA1535 and TA100, while NNP was mutagenic both in the presence and absence of S9 in TA98 (the presence of S9 enhanced the mutagenic yield over the condition without S9). NNP was negative in tester strains TA1537 and WP2 uvrA (pKM101).

Table 1.

Summary of the mutagenicity of N-nitroso propranolol in the Ames test using the preincubation method with different testing conditionsa.

Without S9 Rat liver S9 Hamster liver S9
Tester Strain 10% 30% 10% 30%
30 min 60 min 30 min 60 min 30 min 60 min 30 min 60 min 30 min 60 min
TA98 + + + + + + + + + +
TA100 − − − − − − + + + +
TA1535 − − + − − − + + + +
TA1537 − − − − − − − − − −
WP2 uvrA (pKM101) − − − − − − − − − −
Open in a new tab

−: Not mutagenic.+: Mutagenic: evidence of a dose response with maximum mean revertants per plate is more than 3-fold (for TA1535) or 2-fold (for TA98 and TA100) higher than vehicle controls.

a

Primary data can be found in Supplementary Tables 1–10.

Quantitative analysis of concentration-response Ames data under various NNP testing conditions was conducted by BMD modeling and the results from this quantitative analysis are shown in Table 2. Generally, a lower BMD indicates a more robust mutagenic response (Mittelstaedt et al., 2021; Wills et al., 2016). The magnitude of mutagenic responses for two concentration-responses were considered different if the 90% upper and lower confidence Intervals (CIs) of their BMDs did not overlap. We assessed whether the sensitivity of the Ames test was influenced by 1) induced liver S9 species (rat versus hamster); 2) S9 concentration (10% vs 30%); or 3) preincubation time (30 min vs 60 min). First, in both TA1535 and TA100, hamster liver S9 was more effective than rat liver S9 in activating NNP to a mutagen. Rat liver S9 produced only one positive call (10%, 30 min preincubation in TA1535 with a relatively high benchmark concentration, 36.1 μg/plate of BMD50L). NNP tested negative after metabolism by 30% rat liver S9 in TA1535 and TA100 regardless of preincubation time. In TA98, however, hamster and rat liver S9 produced comparable mutagenic responses. Second, the concentration of liver S9 (10% vs 30%) had no major impact on NNP mutagenicity in TA1535 and TA100. In TA98, however, 10% liver S9 appeared to be more effective than 30% liver S9 regardless of S9 species and preincubation time (Table 2). Finally, the duration of preincubation time did not play a major role in NNP mutagenicity in the Ames test.

Table 2.

Quantitative analysis of concentration-response Ames data of N-nitroso propranolol using the preincubation method with different testing conditionsa.

Strain Liver S9 condition Preincubation (30 min)
 Preincubation (60 min)
BMD50L BMD50U U/L BMD50L BMD50U U/L
TA98 without S9 9.8 24.7 2.5 9.3 15.4 1.7
10% Hamster 5.7 10.1 1.8 6.9 14.5 2.1
10% Rat 7.9 13.0 1.7 7.2 11.5 1.6
30% Hamster 12.8 21.4 1.7 18.3 30.3 1.7
30% Rat 16.2 27.0 1.7 22.4 36.8 1.6
TA100 without S9 Negative Negative
10% Hamster 5.9 21.1 3.6 14.1 34.4 2.4
10% Rat Negative Negative
30% Hamster 7.3 19.9 2.7 24.7 62.2 2.5
30% Rat Negative Negative
TA1535 without S9 Negative Negative
10% Hamster 0.9 3.4 3.8 2.1 6.5 3.1
10% Rat 36.1 108.0 3.0 Negative
30% Hamster 2.3 7.7 3.4 2.5 7.1 2.8
30% Rat Negative Negative
Open in a new tab

BMDL: lower bound of the 90% confidence interval of the BMD (μg/plate).

BMDU: upper bound of the 90% confidence interval of the BMD (μg/plate).

U/L: the ratio of the BMDU to BMDL.

a

Benchmark dose (BMD) analysis (Supplementary Fig. 1) was conducted using web based PROAST software with a critical effect size of 0.5 (i.e., BMD50).

3.2. Genotoxicity in human cells

Three approaches were used to evaluate MN induction by NNP in TK6 cells. Without exogenous metabolism, treatment with up to 100 μM NNP for 24 h did not result in any increase in MN formation over vehicle controls. Higher concentrations were not tested for MN induction due to cytotoxicity. After exogenous bioactivation by hamster liver S9, NNP caused a concentration-dependent increase in %MN frequency after 4-h (2% S9) and 24-h (0.4% S9) treatments (Fig. 2A). For the 4-h treatment with 2% hamster liver S9, 100 μM NNP induced a 4.1-fold increase in % MN over controls. The 24-h continuous incubation with 0.4% hamster liver S9 increased the background level of %MN in TK6 cells but also appeared to increase the sensitivity of the assay. A concentration of 2.5 μM NNP caused a 4.4-fold increase in %MN over controls. NNP also displayed cytotoxicity after metabolic activation by hamster liver S9 in both 4-h and 24-h treatments. For these assays, 100 μM NDEA was used as a positive control and as a reference for nitrosamine-induced genotoxicity. Our results suggested that NNP was more potent than NDEA in inducing MN in TK6 cells after both 4-h and 24-h treatments in the presence of S9. In addition, rat liver S9 also was used for exogenous metabolic activation in this system (data not shown); 24-h continuous incubation with rat liver S9 caused severe cytotoxicity and a high background MN frequency.

Fig. 2.

Fig. 2.

Open in a new tab

N-nitroso propranolol induced genotoxicity in TK6 cells after 4-h and 24-h treatments with hamster liver S9. (A) In vitro micronucleus test. The red line indicates relative cytotoxicity (right y-axis), and the open bars indicate the percentage of micronuclei (left y-axis). (B) In vitro Multiflow assay. Mitomycin C (50 ng/ml) or N-nitrosodiethylamine (NDEA, 100 μM) were used as the positive controls (black bars) without S9 or with hamster S9, respectively. * indicates comparison with the negative control was significant at P < 0.05.

The phosphorylation of H2A.X (γH2A.X) also was measured as a marker for DNA damage. As shown in Fig. 2B, NNP did not induce any increase in γH2A.X fluorescence without S9. At 100 μM, a statistically significant decrease in γH2A.X was detected compared to the vehicle control, which might be attributed to NNP cytotoxicity at this concentration. After 4-h treatment with metabolic activation by 2% hamster S9, NNP induced a concentration-dependent increase in γH2A.X (Fig. 2B). For 24-h treatment with 0.4% hamster S9, 2.5 and 5.0 μM NNP induced increases in γH2A.X. However, the level decreased at the highest concentration tested (10 μM), a concentration causing significant cytotoxicity. We also observed that NNP did not increase the phosphorylation of histone H3, indicating that NNP was not an aneugen (data not shown).

NNP induced concentration-dependent increases in DNA strand breaks evidenced by the Comet assay in both 2D and 3D HepaRG models (Fig. 3). NNP was more cytotoxic in 3D spheroids than in 2D cultured cells (56% vs 82% relative viability at 250 μM) and caused relatively higher % tail DNA in 3D than in 2D cultures (13-fold vs 9-fold increase at the highest concentration compared to control). The lowest concentration that induced positive responses was 25 μM and 50 μM in 3D and 2D cultures, respectively. However, BMD modeling of the DNA damage (% tail DNA) data indicated that the CIs of the two BMD50s for 2D and 3D HepaRG cultures overlapped (data not shown), indicating that the difference in NNP-induced DNA damage between 2D and 3D cell models was not significant.

Fig. 3.

Fig. 3.

Open in a new tab

N-nitroso propranolol induced DNA strand breaks in 2D and 3D human HepaRG cultures. Both 2D (A) and 3D (B) HepaRG cells were exposed to N-nitroso propranolol for 24 h. The relative cell viability (% of control, a measure of cytotoxicity) was measured by the ATP assay (right y-axis and red line). DNA damage (% tail DNA) was detected using the CometChip assay (left y-axis and black bars). The data are expressed as the mean ± SD (n ≥ 3). * indicates comparison with the negative control was significant at P < 0.05.

3.3. TK and HPRT gene mutations in TK6 cells

TK6 cells were treated with NNP in the presence of 1% hamster liver S9 for 4 h and used for gene mutation assays. A preliminary concentration range-finding study indicated that NNP concentrations equal to or above 400 μM produced overt cytotoxicity in TK6 cells (data not shown). In the definitive study, NNP induced an increase in both TK and HPRT gene MFs (Table 3A) and similar results were observed in a repeat experiment (Supplementary Table 12A). Interestingly, the maximum fold-change (about 3–3.5) was found at 100 μM in both independent experiments. The MF induced by NNP decreased from the peak to lower levels at 300 μM, at which concentration NNP was strongly cytotoxic. NDEA at 100 μM was used as a positive control for the mutation assays. The results showed that the mutagenicity of NNP on a molar basis was lower than NDEA in TK6 cells, which was opposite to the observations in the MN assay (Fig. 2A).

Table 3.

TK and HPRT gene mutations induced by 4-h treatment with N-nitroso propranolol (NNP) in TK6 cells with 1% hamster liver S9 activation (A) and CYP2C19-expressing TK6 cells (B).

A (in TK6 cells with 1% hamster liver S9)
 B (in CYP2C19-expressing TK6 cells)
Treatment
 Relative survival (%)
 TK gene MF ( × 10−6)
 HPRT gene MF ( × 10−6)
 Treatment
 Relative survival (%)
 TK gene MF ( × 10−6)
 HPRT gene MF ( × 10−6)
Vehicle controla 100.0 7.3 0.4 Vehicle controla 100.0 6.7 2.0
NNP (μM) 50 115.3 15.5c 4.2c NNP (μM) 0.0625 100.1 13.2 4.5c
100 108.7 21.8c 7.0c 0.125 76.3 18.6c 6.5c
200 86.0 17.8c 3.2c 0.25 75.2 21.3c 19.2c
300 65.9 16.5c 2.6c 0.5 41.8 28.5c 23.0c
Positive control b 66.9 44.2c 23.1c Positive control b 59.9 69.2c 34.1c
Open in a new tab
a

DMSO, 100 μL (the final concentration of 1% v/v).

b

100 μM N-nitrosodiethylamine (A) and 0.1 μg/mL 4NQO (B) were used as positive controls.

c

Indication of positive mutagenic responses (>2-fold over vehicle control).

3.4. Human CYP2C19 accounts for the bioactivation of N-nitroso propranolol

To identify the specific CYPs that account for the bioactivation of NNP, we utilized a panel of eight TK6 cell lines that each endogenously express one different human CYP. The initial concentrations of NNP selected were based on the results from MN assays conducted with exogenous bioactivation by hamster S9 (Fig. 2A). The general hypothesis was that if NNP can be readily transported into the cell, cytotoxicity and genotoxicity should be observed at similar or lower concentrations when metabolism occurs endogenously. As shown in Fig. 4, NNP at concentrations up to 10 μM had no effect on TK6 cells (non-CYP-transduced) and TK6 cells transduced with CYP2A6, CYP2C8, CYP2D6, CYP2E1, or CYP3A4. With TK6 cells transduced with CYP2C19, however, NNP concentrations as low as 0.0625 μM caused a statistically significant increase in %MN. A maximum 5.8-fold increase in %MN, along with a 42% decrease in relative nuclear count (indicating cytotoxicity), was seen at the highest concentration tested (0.5 μM). Higher concentrations could not be tested due to overt cytotoxicity. In addition, CYP1A1 and CYP2C9 had lesser effects on bioactivating NNP. At 10 μM, NNP induced 3.6- and 1.6-fold increases in %MN in CYP1A1-and CYP2C9-expressing TK6 cells, respectively.

Fig. 4.

Fig. 4.

Open in a new tab

N-nitroso propranolol induced micronuclei in TK6 cells transduced with different human CYPs. Cells were treated without exogenous metabolism for 24 h. The red line indicates relative cytotoxicity (right y-axis), and the black bars indicate the percentage of micronucleus formation (left y-axis). The data points represent the means ± SD from at least three independent experiments. * indicates comparison with the negative control was significant at P < 0.05.

The role of CYP2C19 in bioactivating NNP was also shown in human HepG2 cells. We treated human liver HepG2 cells and CYP2C19-expressing HepG2 cells (Chen et al., 2021) with NNP for 24 h and performed the MN assay. NNP did not cause any cytotoxicity or MN formation in HepG2 cells (Fig. 5A). However, NNP induced a concentration-dependent increase in %MN in CYP2C19-expressing HepG2 cells, with a lowest observed adverse effect level (LOAEL) seen at 2 μM (Fig. 5B).

Fig. 5.

Fig. 5.

Open in a new tab

N-nitroso propranolol induced genotoxicity in CYP2C19-expressing HepG2 cells and CYP2C19-expressing TK6 cells. Wild-type HepG2 cells (A) and HepG2 cells transduced with human CYP2C19 (B) were exposed to N-nitroso propranolol without exogenous metabolism for 24 h and grown an additional 16 h (or approximately 1.5 cell cycles). The red line indicates cytotoxicity (right y-axis), and the black bars indicate genotoxicity as percent micronucleus formation (left y-axis). TK6 cells transduced with human CYP2C19 were exposed to N-nitroso propranolol without exogenous metabolism for 4 h (C) or 24 h (D). Nuclear P53 changes (black bars) and γH2A. X formation (gray bars) were determined by the in vitro multiFlow DNA damage assay. All data points represent the means ± SD from at least three independent experiments. * indicates comparison with the negative control was significant at P < 0.05.

3.5. N-nitroso propranolol induces DSBs and mutations in CYP2C19-expressing TK6 cells

To further examine the genotoxicity of NNP after CYP2C19 bioactivation, we conducted the MultiFlow DNA damage assay to measure γH2A.X formation and nuclear-localized P53 content, which are biomarkers of DSBs. As shown in Fig. 5, NNP produced a concentration-dependent increase in γH2A.X in CYP2C19-expressing TK6 cells after both 4- and 24-h exposures, beginning with the lowest concentration tested (0.625 μM). A 24-h treatment with 0.5 μM NNP increased the γH2A.X fluorescence signal more than 3-fold over the controls. On the other hand, NNP only caused increases in nuclear P53 after 24 h, suggesting that P53 was stabilized upon DNA damage, leading to its accumulation.

Finally, we observed that NNP was mutagenic in TK6 cells transduced with human CYP2C19 (Table 3B) and then confirmed these results by a repeat experiment (Supplementary Table 12B). Relative survival (RS) was determined by cloning efficiency immediately following treatment (CE0). NNP at 0.5 μM decreased the RS in CYP2C19-expressing TK6 cells to 41.8% and 39.4% in two independent experiments. NNP induced concentration-dependent increases in both TK and HPRT MFs. The highest concentration of NNP tested (0.5 μM) led to 4.3- and 11.5-fold MF increases over control in the TK and HPRT gene, respectively (Table 3B). Notably, no exogenous metabolic activation systems or cofactors were included in the treatments, indicating that the cytotoxicity and mutagenicity of NNP were manifested solely by endogenous CYP2C19 expression.

4. Discussion

The interest in nitrosamines as drug impurities has increased as the scope of the number of drugs impacted by these impurities has increased (Schlingemann et al., 2023). While the mutagenicity and carcinogenicity of some nitroso dialkylamines have been characterized extensively, relatively limited data exist on the genotoxicity of many nitrosamines, especially NDSRIs. NNP, the nitrosamine derived from propranolol, is one of a handful of NDSRIs that have both published bacterial mutagenicity data and mammalian cell genotoxicity data (Brambilla and Martelli, 2007). These data, however, are inconsistent: NNP previously was reported to be negative in the Salmonella typhimurium gene mutation test and strongly genotoxic in both rat and human hepatocytes.

While it is not rare to find inconsistent genotoxicity findings for a compound evaluated in multiple studies over a period of 40 years, we were cautious about accepting the historical negative Ames test results on NNP without further examination. Importantly, the two studies that showed NNP was not mutagenic (Raisfeld-Danse and Chen, 1983; Zhang et al., 1983) were conducted using only rat liver S9 as the exogenous metabolic activation system. In the same year that the Ames NNP findings were reported, a study testing more than 150 nitrosamines demonstrated that a number of carcinogenic nitroso compounds that were not activated to mutagens in the Ames test by rat liver microsomal fraction were mutagenic using a hamster liver microsomal fraction (Lijinsky and Andrews, 1983). Since this study was conducted using only the TA1535 tester strain and using methods pre-dating current OECD TG 471 test guideline, a recent retrospective analysis argued that Ames tests conducted with hamster and rat liver S9s as the exogenous metabolism system had similar sensitivities for detecting carcinogenic nitrosamines when assays were conducted using all five tester strains according to the OECD TG 471 (Trejo-Martin et al., 2022). Even so, several carcinogenic nitroso compounds that tested positive with hamster liver S9 in OECD-compliant Ames tests showed negative results when using rat liver S9 in the head-to-head comparisons. Moreover, in many cases (including our unpublished data), hamster liver S9 appears to induce a stronger revertant colony response when compared to rat liver S9.

In the current study, we found that rat liver S9 was much less effective than hamster liver S9 in bioactivating NNP to a mutagen in TA1535 and TA100 (Tables 1 and 2), both of which revert the same base pair substitution mutation (hisG46). The Ames test is considered a hazard identification assay and there is no quantitative correlation implied from these Ames findings for the relative tumorigenicity of a nitrosamine in rats and hamsters. While it is possible that increased sensitivity may lead to decreased specificity, in our case, the positive mutagenic response of NNP in the Ames test is concordant with its positive clastogenic and mutagenic responses in human cells. This includes human cell assays employing endogenous human cell metabolic activation systems. Therefore, the weight of evidence suggests that the bacterial mutagenicity of NNP indicates a potential risk for exposed humans.

In addition to the mutagenicity findings in TA1535 and TA100, we found that rat and hamster liver S9 were similar in activating NNP to a mutagen in TA98 (Table 2). Therefore, the species of the S9 did not totally explain the difference between our results and the results of the previous two Ames tests which indicated NNP was non-mutagenic in TA98 (Raisfeld-Danse and Chen, 1983; Zhang et al., 1983). We postulate that the strong positive response observed in our study for TA98 may be attributed to two reasons. First, the S9 concentrations used in the previous two Ames studies were not specified. Nitrosamines, even potent carcinogenic nitrosamines such as NDMA and NDEA, are often positive in the Ames test only by using high concentrations of liver S9 (Ishidate et al., 1988). Therefore, we used relatively high levels of rodent liver S9 (10% and 30% v/v), which could have led to increased sensitivity. Second, although the duration of the preincubation time did not play a critical role in the sensitivity of the assay, performing the preincubation version of the Ames test in our study may have resulted in increased sensitivity to the mutagenicity of NNP relative to the plate incorporation strategy used in the two previous studies. Preincubation conducted with exogenous metabolic activation systems has been recommended to improve the efficiency of detecting the mutagenicity of nitrosamines in bacterial mutation assays (Gatehouse et al., 1994; OECD, 2020).

Interestingly, NNP was positive in TA98 even without metabolic activation, although the mutagenic response was weaker than that in the presence of rat or hamster liver S9s (Supplementary Tables 5 and 6, &11). TA98 detects frameshift mutations at a run of alternating GC base pairs in hisD3052 and historical data showed that TA98 was relatively ineffective in detecting the mutagenicity of nitroso dialkylamines (Yahagi et al., 1977). We excluded a possible solvent effect by comparing the mutagenicity of NNP dissolved in acetone, methanol, or DMSO (Supplementary Table 11), which showed NNP produced very similar mutational responses in TA98 using three different vehicles. Although not totally unprecedented, it is uncommon for an N-nitrosamine to produce a mutagenic response in the Ames test without exogenous metabolism. Such a response might be a result of an N-nitroso compound breaking down either spontaneously or under the conditions of the assay. N-nitroso chlordiazepoxide previously was shown to induce concentration-dependent HPRT mutations in hamster V79 cells without exogenous activation (Brambilla et al., 1989). This finding for NNP highlights the diverse characteristics of NDSRIs, but genotoxic effects in the presence of S9 is still a common characteristic of carcinogenic nitrosamines.

Previous studies also indicate that a standard rat liver S9 mix was relatively ineffective at activating nitrosamines to mutagens in assays conducted with human cells. For example, 2% Aroclor-1254-induced rat liver S9 failed to biotransform NDEA to produce any genotoxic or cytotoxic responses in TK6 cells (Bernacki et al., 2016). In another study, up to 10 mM NDEA did not induce MN formation in TK6 cells using bioactivation with 1% rat liver S9 (Liviac et al., 2011). In the current study, we showed that relatively low concentrations of both NNP and NDEA induced MN with hamster liver S9 activation (Fig. 2A). The modified treatment regimen, particularly the replacement of rat with hamster liver S9 and the inclusion of calcium as a cofactor, likely improved the detection sensitivity for nitrosamines in TK6 cells. A similar hamster liver S9 mixture also was effective in detecting the mutagenicity of NNP and NDEA in TK6 cells (Table 3A). For NDSRIs, there are few mutagenicity data in mammalian cells, particularly in human cells. Our results showed a concordance between mutagenicity in bacteria and human cells, which at least partially addresses the question of whether the mutagenic effects of NNP observed in bacteria can potentially be found in mammalian systems.

In addition, a previous study (Robbiano et al., 1991) and our data demonstrate that metabolism is required for NNP to exert its genotoxicity in human cells. This suggests that the direct-acting mutagenicity of NNP in the Ames test may be limited to this assay, and thus may be mediated by bacterial metabolism. With regard to the specific CYP isoforms involved in the activation, CYP2A6 and CYP2E1 previously have been shown to be critical for activating small-to medium-sized N-alkyl nitrosamines in both bacteria and human cells (Kamataki et al., 2002; Li et al., 2022). However, it is predicted that, as the size of nitroso compounds increases, and depending on the charge status of each nitrosamine, other enzymes such as CYP2C9, CYP2D6, CYP2C19, and CYP3A4 will become more important for nitrosamine activation (Cross and Ponting, 2021). Identifying the specific CYP enzymes that account for the bioactivation of NNP helps better understand the mechanism of NNP genotoxicity, as well as potentially explaining differences in observations between assays conducted on NNP genotoxicity and identifying a potential root cause for inter-individual differences in responses to NNP exposure.

In our study, we demonstrated that CYP2C19 is the main human enzyme accounting for the bioactivation of NNP (Figs. 4 and 5). Propranolol, the parent compound of NNP, is metabolized by CYP2D6 to form 4- and 5-hydroxy-propranolol, which are the active β-adrenergic receptor antagonists (Masubuchi et al., 1994). CYP2D6 may preferably bind to substrates with cationic sites (Lewis and Ito, 2010). Theoretically, the nitrosation reaction on the secondary –NH– group of propranolol decreases its positive charge (Fig. 1), making NNP a more neutral compound that can be metabolized by CYP2C19. Indeed, we showed that relatively low concentrations of NNP produced significant, concentration-related increases in genotoxicity and mutagenicity in CYP2C19-expressing TK6 cells. Notably, the basal level of CYP2C19 enzymatic activity also was high in HepaRG cells (Berger et al., 2016), which may have been responsible for the NNP-induced DNA damage in this cell model (Fig. 3).

In summary, we revisited the issue of NNP mutagenicity and genotoxicity using multiple testing approaches. This study demonstrated that NNP is positive for bacterial mutation in the Ames test and that hamster liver S9 is generally more effective than rat liver S9 in detecting the mutagenicity of NNP in bacteria. NNP also damages DNA, induces MN, and is mutagenic in human cells; and CYP2C19 appears to be the main human enzyme accounting for its bioactivation. Our results provide novel information for the hazard identification and risk assessment of NNP, and the types of experiments with expanded testing conditions utilized in this study may also be useful for exploring and characterizing the mutagenicity and genotoxicity of other nitrosamines and NDSRIs.

Supplementary Material

12 Supplementary Table & 1 Figureble

Acknowledgments

This article reflects the views of the authors and should not be construed to represent the views or policies of the U.S. Food and Drug Administration (FDA). Any mention of commercial products, their sources, or their use in connection with material reported herein is for clarification only and is not to be construed as either an actual or implied endorsement of such products. We thank Drs. Tao Chen and Dayton Petibone for their critical review of this manuscript.

Funding body information

This work was partly supported by funding from the FDA Center for Drug Evaluation and Research (CDER) Regulatory Science Research program. YL was supported by an appointment to the Postgraduate Research Program at the FDA National Center for Toxicological Research (NCTR) administered by the Oak Ridge Institute for Science Education through an interagency agreement between the U.S. Department of Energy and the U.S. FDA.

Footnotes

CRediT authorship contribution statement

Xilin Li: Methodology, Investigation, Validation, Writing – original draft. Yuan Le: Investigation, Validation. Ji-Eun Seo: Investigation, Validation. Xiaoqing Guo: Investigation, Validation. Yuxi Li: Investigation, Validation. Si Chen: Investigation, Validation. Roberta A. Mittelstaedt: Investigation, Validation. Nyosha Moore: Investigation, Validation. Sharon Guerrero: Investigation, Validation. Audrey Sims: Investigation, Validation. Sruthi T. King: Conceptualization, Writing – review & editing. Aisar H. Atrakchi: Conceptualization, Funding acquisition, Project administration, Writing – review & editing. Timothy J. McGovern: Conceptualization, Writing – review & editing. Karen L. Davis-Bruno: Conceptualization, Writing – review & editing. David A. Keire: Conceptualization, Writing – review & editing, Project administration, Funding acquisition. Rosalie K. Elespuru: Conceptualization, Writing – review & editing. Robert H. Heflich: Methodology, Resources, Conceptualization, Writing – review & editing, Supervision, Project administration, Funding acquisition. Nan Mei: Methodology, Resources, Conceptualization, Writing – review & editing, Supervision, Project administration, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.yrtph.2023.105410.

Data availability

Data will be made available on request.

References

  1. Avlasevich S, et al. , 2021. Biomarkers of DNA damage response improve in vitro micronucleus assays by revealing genotoxic mode of action and reducing the occurrence of irrelevant positive results. Mutagenesis 36, 407–418. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Barton AL, et al. , 2015. Efficacy and safety of high-dose propranolol for the management of infant supraventricular tachyarrhythmias. J. Pediatr 166, 115–118 [DOI] [PubMed] [Google Scholar]
  3. Bellec G, et al. , 1996. Cytochrome P450 metabolic dealkylation of nine N-nitrosodialkylamines by human liver microsomes. Carcinogenesis 17, 2029–2034 [DOI] [PubMed] [Google Scholar]
  4. Berger B, et al. , 2016. Comparison of liver cell models using the basel phenotyping cocktail. Front. Pharmacol 7, 443. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bernacki DT, et al. , 2016. γH2AX and p53 responses in TK6 cells discriminate promutagens and nongenotoxicants in the presence of rat liver S9. Environ. Mol. Mutagen 57, 546–558. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Brambilla G, Martelli A, 2007. Genotoxic and carcinogenic risk to humans of drug-nitrite interaction products. Mutat. Res 635, 17–52. [DOI] [PubMed] [Google Scholar]
  7. Brambilla G, et al. , 1989. Genotoxicity of N-nitrosochlordiazepoxide in cultured mammalian cells. Toxicol. Appl. Pharmacol 97, 480–488. [DOI] [PubMed] [Google Scholar]
  8. Bryce SM, et al. , 2007. In vitro micronucleus assay scored by flow cytometry provides a comprehensive evaluation of cytogenetic damage and cytotoxicity. Mutat. Res 630, 78–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bryce SM, et al. , 2016. Genotoxic mode of action predictions from a multiplexed flow cytometric assay and a machine learning approach. Environ. Mol. Mutagen 57, 171–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Callander RD, et al. , 1995. Evaluation of phenobarbital/beta-naphthoflavone as an alternative S9-induction regime to Aroclor 1254 in the rat for use in in vitro genotoxicity assays. Mutagenesis 10, 517–522. [DOI] [PubMed] [Google Scholar]
  11. Canada, 2022. Health product recall. Government of Canada. https://recalls-rappels.canada.ca/en/alert-recall/inderal-nitrosamine-impurity. (Accessed 2 May 2023). [Google Scholar]
  12. Chen S, et al. , 2021. Characterization of cytochrome P450s (CYP)-overexpressing HepG2 cells for assessing drug and chemical-induced liver toxicity. J. Environ. Sci. Health C Toxicol. Carcinog 39, 68–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Chen T, et al. , 2022. Thymidine kinase(+/−) mammalian cell mutagenicity assays for assessment of nanomaterials. Front Toxicol. 4, 864753. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Cross KP, Ponting DJ, 2021. Developing structure-activity relationships for N-nitrosamine activity. Comput. Toxicol 20, 100186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. FDA, 2018. M7(R1) assessment and control of DNA reactive (mutagenic) impurities in pharmaceuticals to limit potential carcinogenic risk. https://www.fda.gov/media/85885/download. (Accessed 2 May 2023).
  16. Fujita K, Kamataki T, 2001. Role of human cytochrome P450 (CYP) in the metabolic activation of N-alkylnitrosamines: application of genetically engineered Salmonella typhimurium YG7108 expressing each form of CYP together with human NADPH-cytochrome P450 reductase. Mutat. Res 483, 35–41. [DOI] [PubMed] [Google Scholar]
  17. Gatehouse D, et al. , 1994. Recommendations for the performance of bacterial mutation assays. Mutat. Res 312, 217–233. [DOI] [PubMed] [Google Scholar]
  18. ICH, 2017. Assessment and control of DNA reactive (mutagenic) impurities in pharmaceuticals to limit potential carcinogenic risk M7(R1). https://database.ich.org/sites/default/files/M7_R1_Guideline.pdf. (Accessed 2 May 2023).
  19. Ishidate M Jr., et al. , 1988. A comparative analysis of data on the clastogenicity of 951 chemical substances tested in mammalian cell cultures. Mutat. Res 195, 151–213. [DOI] [PubMed] [Google Scholar]
  20. Kalam MN, et al. , 2020. Clinical pharmacokinetics of propranolot hydrochloride: a review. Curr. Drug Metabol 21, 89–105. [DOI] [PubMed] [Google Scholar]
  21. Kamataki T, et al. , 2002. Role of human cytochrome P450 (CYP) in the metabolic activation of nitrosamine derivatives: application of genetically engineered Salmonella expressing human CYP. Drug Metab. Rev 34, 667–676. [DOI] [PubMed] [Google Scholar]
  22. Kikugawa K, et al. , 1987. Formation of a higly mutagenic diazo compound from the bamethan-nitrite reaction. Mutat. Res. Fund Mol. Mech. Mutagen 177, 35–43. [DOI] [PubMed] [Google Scholar]
  23. Lewis DF, Ito Y, 2010. Human CYPs involved in drug metabolism: structures, substrates and binding affinities. Expet Opin. Drug Metabol. Toxicol 6, 661–674. [DOI] [PubMed] [Google Scholar]
  24. Li X, et al. , 2020a. Development and application of TK6-derived cells expressing human cytochrome P450s for genotoxicity testing. Toxicol. Sci 175, 251–265. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Li X, et al. , 2020b. Evaluation of pyrrolizidine alkaloid-induced genotoxicity using metabolically competent TK6 cell lines. Food Chem. Toxicol 145, 111662. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Li X, et al. , 2022. Genotoxicity evaluation of nitrosamine impurities using human TK6 cells transduced with cytochrome P450s. Arch. Toxicol 96, 3077–3089. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lijinsky W, Andrews AW, 1983. The superiority of hamster liver microsomal fraction for activating nitrosamines to mutagens in Salmonella typhimurium. Mutat. Res 111, 135–144. [DOI] [PubMed] [Google Scholar]
  28. Lijinsky W, Epstein SS, 1970. Nitrosamines as environmental carcinogens. Nature 225, 21–23. [DOI] [PubMed] [Google Scholar]
  29. Liviac D, et al. , 2011. Genotoxic evaluation of the non-halogenated disinfection by-products nitrosodimethylamine and nitrosodiethylamine. J. Hazard Mater 185, 613–618. [DOI] [PubMed] [Google Scholar]
  30. Magee PN, Barnes JM, 1967. Carcinogenic nitroso compounds. Adv. Cancer Res 10, 163–246. [DOI] [PubMed] [Google Scholar]
  31. Martelli A, et al. , 1994. Low clastogenic activity in vivo of the N-nitroso derivatives of 5 beta-adrenergic-blocking drugs proved to be potent genotoxins in vitro. Toxicol. Lett 73, 185–191. [DOI] [PubMed] [Google Scholar]
  32. Masubuchi Y, et al. , 1994. Cytochrome P450 isozymes involved in propranolol metabolism in human liver microsomes. The role of CYP2D6 as ring-hydroxylase and CYP1A2 as N-desisopropylase. Drug Metab. Dispos 22, 909–915. [PubMed] [Google Scholar]
  33. Mittelstaedt RA, et al. , 2021. Differentiating between micronucleus dose-responses induced by whole cigarette smoke solutions with Benchmark Dose potency ranking. Mutat. Res. Genet. Toxicol. Environ. Mutagen 866, 503351. [DOI] [PubMed] [Google Scholar]
  34. O’Neill JP, et al. , 1982. Cytotoxicity and mutagenicity of dimethylnitrosamine in mammalian cells (CHO/HGPRT system); enhancement by calcium phosphate. Environ. Mutagen 4, 7–18. [DOI] [PubMed] [Google Scholar]
  35. OECD, 2016. OECD guideline for the testing of chemicals, section 4. https://www.oecd-ilibrary.org/environment/oecd-guidelines-for-the-testing-of-chemicals-section-4-health-effects_20745788. (Accessed 2 May 2023).
  36. OECD, 2020. Bacterial reverse mutation test, OECD test guideline No. 471. https://www.oecd-ilibrary.org/environment/test-no-471-bacterial-reverse-mutation-test_9789264071247-en. (Accessed 2 May 2023).
  37. Ozhan G, Alpertunga B, 2003. Genotoxic activities of drug-nitrite interaction products. Drug Chem. Toxicol 26, 295–308. [DOI] [PubMed] [Google Scholar]
  38. Preussmann R, 1984. Occurrence and exposure to N-nitroso compounds and precursors. IARC Sci. Publ 3–15. [PubMed] [Google Scholar]
  39. Raisfeld-Danse IH, Chen J, 1983. Drug interactions. III. Formation of nitrosamines from therapeutic drugs. Formation, mutagenic properties and safety assessment of propranolol hydrochloride with respect to the intragastric formation of N-nitrosopropranolol under conditions found in patients. J. Pharmacol. Exp. Therapeut 225, 713–719. [PubMed] [Google Scholar]
  40. Robbiano L, et al. , 1991. Formation of the N-nitroso derivatives of six beta-adrenergic-blocking agents and their genotoxic effects in rat and human hepatocytes. Cancer Res. 51, 2273–2279. [PubMed] [Google Scholar]
  41. Schlingemann J, et al. , 2023. The landscape of potential small and drug substance related nitrosamines in pharmaceuticals. J. Pharmaceut. Sci 112, 1287–1304. [DOI] [PubMed] [Google Scholar]
  42. Seo JE, et al. , 2022. Evaluation of an in vitro three-dimensional HepaRG spheroid model for genotoxicity testing using the high-throughput CometChip platform. ALTEX 39, 583–604. [DOI] [PubMed] [Google Scholar]
  43. Seo JE, et al. , 2019. Quantitative comparison of in vitro genotoxicity between metabolically competent HepaRG cells and HepG2 cells using the high-throughput high-content CometChip assay. Arch. Toxicol 93, 1433–1448. [DOI] [PubMed] [Google Scholar]
  44. Sluggett GW, et al. , 2018. Artifactual degradation of secondary amine-containing drugs during accelerated stability testing when saturated sodium nitrite solutions are used for humidity control. J. Pharm. Biomed. Anal 149, 206–213. [DOI] [PubMed] [Google Scholar]
  45. Trejo-Martin A, et al. , 2022. Use of the bacterial reverse mutation assay to predict carcinogenicity of N-nitrosamines. Regul. Toxicol. Pharmacol 135, 105247. [DOI] [PubMed] [Google Scholar]
  46. Wills JW, et al. , 2016. Empirical analysis of BMD metrics in genetic toxicology part I: in vitro analyses to provide robust potency rankings and support MOA determinations. Mutagenesis 31, 255–263. [DOI] [PubMed] [Google Scholar]
  47. Yahagi T, et al. , 1977. Mutagenicities of N-nitrosamines on Salmonella. Mutat. Res 48, 121–129. [DOI] [PubMed] [Google Scholar]
  48. Yamazaki H, et al. , 1992. Cytochrome P450 2E1 and 2A6 enzymes as major catalysts for metabolic activation of N-nitrosodialkylamines and tobacco-related nitrosamines in human liver microsomes. Carcinogenesis 13, 1789–1794. [DOI] [PubMed] [Google Scholar]
  49. Zhang S, et al. , 1983. Derivatives of beta-adrenergic antagonists. N-nitrosopropranolol and N-hydroxypropranolol and its aldonitrone. J. Med. Chem 26, 455–458. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

12 Supplementary Table & 1 Figureble
NIHMS2019427-supplement-12_Supplementary_Table___1_Figureble.docx (428.3KB, docx)

Data Availability Statement

Data will be made available on request.

RESOURCES